The program is built for the gap beneath a traditional bank's minimum: too small for a commercial loan officer to bother with, too specific for a credit card to cover cleanly. It has existed for decades, but the SBA re-promoted it in the spring of 2026, and it remains one of the more overlooked tools in the agency's lineup.

What is an SBA microloan, exactly?

It is capital the SBA does not lend directly. The agency provides funds to a network of nonprofit, community-based organizations called intermediaries, and those intermediaries make the credit decisions, set the exact rate within the federal range, and disburse the money. Loan size runs from a few hundred dollars up to the $50,000 ceiling.

Matt Coleman, the SBA's Atlantic Regional Administrator, put the pitch plainly: "Small businesses are the backbone of local economies, yet too many entrepreneurs overlook one of the most accessible financing tools available." Intermediaries are also required to offer guidance around the loan — before, during and after the money lands — which is part of what separates the program from a straight commercial loan.

What can the money actually be used for?

Eligible uses are working capital, inventory, supplies, furniture, fixtures, and machinery or equipment — the operating costs of running or expanding a business. The restriction is just as specific: microloan funds cannot be used to buy real estate or to pay off existing debts. A founder trying to refinance a merchant cash advance, or buy the building a shop operates out of, is looking at the wrong program.

Who qualifies, and who administers the loan?

Eligibility follows standard SBA rules: the business must operate for profit, be based in the United States, and fall within the agency's small-business size standards. Certain nonprofit childcare centers are also eligible, an exception carved out specifically in the program's rules. Beyond that baseline, the real gatekeeping happens at the intermediary level — each nonprofit lender sets its own underwriting standard, which is why approval odds and required documentation vary by lender rather than by a single national bar.

What does it cost, and how fast does it get repaid?

Rates run between 8% and 13%, with the exact number set by the intermediary rather than the SBA itself. Repayment terms run up to seven years, though the actual term depends on the loan amount and its intended use — a $5,000 inventory loan and a $45,000 equipment loan are not going to carry the same schedule. Because the SBA does not publish a single fixed rate, comparing offers across two or more intermediaries before signing is the only way to know if a given rate is competitive.

How does it compare to other small-business financing?

The clearest contrast is with conventional bank financing, which can run terms as long as 25 years but comes with stricter credit and revenue requirements. Microloans trade that scale for accessibility: smaller amounts, more lenient qualification standards, and a workable option for owners with thinner credit histories or businesses too new to show a multi-year track record. The SBA is not the only source — the USDA's Farm Service Agency runs a microloan program for agricultural businesses, and nonprofit peer-to-peer platforms serve similar borrowers outside the federal system — but the SBA's version is the one most general small businesses will encounter first.

How does a founder actually apply?

There is no single national application, because there is no single national lender. A founder starts by finding an SBA-approved intermediary that operates in their area or industry, since each one sets its own paperwork, timeline, and underwriting bar rather than following one federal form. That local structure is also the point: the same intermediary reviewing the loan is typically the one offering the ongoing guidance the program is built around, before and after the money arrives.

Because terms and documentation differ by lender, the practical first step is comparison — checking rate, required collateral, and the intermediary's own track record with businesses of a similar size and stage before signing anything. A microloan approved quickly by one intermediary is not automatically the best offer available; it is simply the first one found.

The bottom line for founders

A microloan will not fund a buildout or retire old debt, and at a $50,000 ceiling it is not built to scale a company past its early stage. What it is built for is the specific, unglamorous gap — the first inventory order, the used equipment, the working capital to cover payroll through a slow month — where a bank will not return the call and a nonprofit lender will.

That curve is the single most useful number a first-time founder can carry into year one, and it is routinely mangled. It is not the widely repeated claim that nine in ten businesses fail. It is not a promise either. It is a measured attrition rate for a cohort of real establishments, tracked one year at a time by a federal statistical agency, and it has looked broadly the same for three decades.

What share of new businesses actually survive five years?

Roughly half. The Bureau of Labor Statistics survival table follows each birth cohort of private-sector establishments from its opening quarter forward, reporting how many are still operating at each anniversary. For the March 2013 cohort, survival since birth ran 79.6 percent at one year, 50.6 percent at five years and 34.7 percent at ten.

Two things follow from that shape. The steepest single drop happens in the first twelve months, when roughly one establishment in five closes. After that the curve flattens: the business that clears year one is statistically a different proposition from the business that opened last quarter.

The agency also publishes a second rate alongside it — survival of the previous year's survivors — which measures the annual hazard rather than the cumulative one. That distinction matters, because a founder in year four is not facing the same odds as a founder in month four.

Has the survival curve changed over time?

Not dramatically. The BLS charts cohorts by opening year from 1994 through 2015 and finds that survival follows a similar path regardless of birth year. Measured a year after opening, the 1994 cohort stood at 79.6 percent, the 2000 cohort at 78.4 percent and the 2008 cohort — the businesses that opened directly into the financial crisis — at 75.2 percent.

A few points of difference in a downturn is not nothing, but it is smaller than the folklore suggests. Macroeconomic timing shifted the first-year number by about four percentage points for the worst-timed cohort in the series. The longer tail is where the real erosion happens: the 1994 cohort was down to 19.5 percent twenty-two years after opening, and the 2000 cohort to 26.3 percent by year seventeen.

Survival rates also vary by industry, which the BLS reports separately. A national average is a blend of sectors with very different capital requirements, lease obligations and customer-acquisition costs, and no founder operates in the average.

How big is the population these rates describe?

Very large, which is why small differences in the rate translate into enormous absolute numbers. The SBA Office of Advocacy counted 33,185,550 small businesses in the United States, or 99.9 percent of American businesses, employing 61.7 million people — 46.4 percent of private-sector employees.

Those firms paid 39.4 percent of private-sector payroll and generated 32.6 percent of known export value, the agency reported. Between 1995 and 2021 they created 17.3 million net new jobs, 62.7 percent of net jobs created over that period.

Set against a population that size, a five-year survival rate near 50 percent describes millions of closures and millions of survivors simultaneously. Both halves of that sentence are true, and most retellings of the statistic keep only the half that suits the argument.

What does business formation data add?

It captures the front end of the funnel that survival rates measure at the back. The Census Bureau's Business Formation Statistics provide what the bureau describes as timely, high-frequency information on new business applications and formations in the United States, built with economists from the Federal Reserve Board, the Federal Reserve Bank of Atlanta, the University of Maryland and the University of Notre Dame.

Applications and survival are separate measurements of separate things. A quarter of heavy application volume tells a founder about the competitive environment they are entering; the survival table tells them what happens to the entrants several years later. Reading either one alone produces a distorted picture of how crowded or how forgiving a market is.

What the numbers can and cannot tell a founder

They describe a population, not a business. Survival statistics are a base rate: the starting point before anything specific about a company — its margins, its concentration of customers, its founder's prior operating experience — is taken into account. They cannot forecast a single outcome, and no cohort rate should be read as a prediction about any one venture.

What they usefully correct is the framing. The commonly cited idea that the overwhelming majority of new businesses vanish quickly is not what the federal series shows; the first year is the dangerous one, and the curve then flattens into a long, slow decline. Planning cash runway around a brutal first twelve months, rather than around a mythical universal failure rate, is closer to what the record supports.

This is information drawn from published federal statistics, not financial advice. Decisions about financing, leases or hiring turn on facts specific to a business, and the survival table is not a substitute for them.

The lesson the data actually supports is narrow and worth stating plainly: the first year carries the largest share of the risk, and clearing it changes the odds materially. Everything after that is a slower attrition that rewards businesses able to keep operating rather than businesses that opened at a lucky moment.

Business News 7 publishes information, not investment advice; funding coverage is market data, not recommendations.

What Were the Rounds?

Blue Origin's $10 billion raise — the space company's first external funding — topped the month, followed by Safe Superintelligence's reported $5 billion from Nvidia, Moonshot AI at $3.5 billion, and Kling AI at $2.8 billion. Germany contributed billion-dollar rounds for defense-tech company Helsing and drone-maker Quantum Systems. The month's total for mega-rounds set the count record but not the dollar record, and AI companies absorbed about $35 billion — roughly 53 percent of the month's funding, a lower share than Q1's 80 percent peak. U.S. startups took $39 billion, about 59 percent.

What Did Exits Look Like?

The exit side moved too: venture-backed M&A topped $9 billion in July with five exits above $1 billion, including Nscale's roughly $1.65 billion purchase of Anyscale and Cyera's $1 billion acquisition of Oasis Security. Twelve venture-backed IPOs debuted valued above $1 billion, led by ChangXin Memory Technologies at roughly an $85 billion debut — up 466 percent — and Bending Spoons at $18.5 billion. For founders, that combination — record mega-round counts plus an open acquisition and IPO window — is the strongest exit environment signal since 2021, with the caveat that both buyer types are concentrating on AI and adjacent infrastructure.

What Does It Change for Founders?

Two practical reads. The record round count with a falling AI share suggests capital is broadening within technology — defense, space, quantum, and memory hardware all posted billion-dollar months — which is encouragement for hard-tech and deep-tech founders whose theses sat outside last year's narrow AI window. And the healthy M&A and IPO activity changes negotiation posture: acquirers and public-market comparables are both active, so founders fielding offers in the second half can benchmark against real marks rather than theoretical ones. The standing caution from the Q1 data still applies: these records describe a market of mega-rounds most startups never touch, but the direction — more rounds, more acquirers, more listings — improves the climate at every stage beneath the headlines.

The takeaway for founders: July's fourteen billion-dollar rounds and $9 billion-plus of venture M&A mark the most open capital-and-exit environment of the cycle — hard-tech founders in particular are raising money in a market newly willing to price their work.

Business News 7 publishes information, not investment advice; funding coverage is market data, not recommendations.

What Else Was in the Week?

Oratomic, a South Pasadena startup building neutral-atom quantum hardware, raised a $300 million Series A co-led by Arch Venture Partners, Khosla Ventures, and Spark Capital across a sixteen-investor syndicate including Bezos Expeditions — one of the largest quantum Series A rounds on record. Below it: Quaise Energy's $134 million Series B for millimeter-wave geothermal drilling, Prime Intellect's $130 million Series A for distributed AI compute, Gauntlet's $125 million from SBI Group for DeFi risk management, Norm AI's $120 million Series C at a reported $1.2 billion valuation for translating laws into compliance software, Venus Aerospace's $91 million for hypersonic propulsion, EDX Markets' $76 million, and Fore Biotherapeutics' $67.4 million precision-oncology round.

What Does It Change for Founders?

The week's two billion-dollar rounds landed in different capital systems — SambaNova in growth venture, Keyfactor in private equity — and that pairing is the signal worth reading. Private equity's willingness to write billion-dollar checks to mature software companies deepens the exit-and-funding menu for founders: PE is now a realistic alternative to an IPO or a strategic sale for profitable, boring-on-purpose infrastructure businesses, and a founder deciding whether to raise another venture round can put a PE recapitalization on the same whiteboard. SambaNova's round, meanwhile, extends the year's hardware-inference pattern: investors keep paying premium valuations for the silicon and systems that run AI in production, with an $11 billion post-money on roughly $2.5 billion previously raised. And Oratomic's $300 million Series A sets the early-stage benchmark for quantum — a sector where the same dynamic that played out in AI (infrastructure first, applications later) is beginning to price in.

The takeaway for founders: a week with twin $1 billion rounds — one venture, one private equity — shows both capital systems bidding hard for infrastructure assets at once; founders in security, compute, and deep tech should be talking to both.

Business News 7 publishes information, not investment advice; funding coverage is market data, not recommendations.

What Was in the Week?

Beyond the two billion-dollar deals, Groq — the AI inference cloud provider — raised $650 million as it scales capacity, marking the week's clearest signal: two of the top three rounds went to companies whose product is running other companies' AI models. Ollin Biosciences raised $330 million for vision-disease therapies, and General Intuition — a foundational AI model trained through gameplay — raised a $320 million Series A at a $2.3 billion valuation. Peregrine Technologies took $250 million for its public-safety software platform at a $6.8 billion valuation, with Quantifind and the frontier lab Mirendil at $200 million each, and Upscale AI and Osanni Bio at $190 million apiece. European rounds added Berlin defense-tech company Stark at $569 million and Paris health insurer Alan's €480 million.

What Does It Change for Founders?

The week's composition confirms a trend visible since Q1: the inference layer — the infrastructure that runs AI models in production — has become venture capital's favored theme within the AI stack, distinct from the model developers themselves. Baseten and Groq selling compute and tooling to companies integrating AI mirrors what Supabase's raise signaled three weeks earlier: investors are funding the merchants of the gold rush at richer multiples than most miners. The second signal is valuation discipline by category: infrastructure rounds priced at $11-13 billion and platform rounds like Peregrine's at $6.8 billion, while a Series A with a credible AI thesis — General Intuition at $2.3 billion, Mirendil's $200 million seed — shows the very top of the early-stage market still commands 2021-era pricing. For founders raising below that tier, the lesson is positioning: the weeks keep proving that "we make AI work in production" out-raises "we are building AI" — a framing decision made in the pitch, not the technology.

The takeaway for founders: a $4.5 billion week with inference infrastructure on top repeats the year's clearest pattern — the money follows whoever makes AI cheaper and faster to run, and founders selling into that layer are pricing their rounds accordingly.

Business News 7 publishes information, not investment advice; funding coverage is market data, not recommendations.

What Else Was in the Week?

The $500 million tier crowded quickly. Impulse Space raised a $500 million Series D for spacecraft propulsion, passing $1 billion raised to date. Supabase, the open-source platform developers use to build AI applications, raised $500 million led by GIC at a $10.5 billion valuation — a signal of where AI-era infrastructure spending flows. Flourish, a foundational-AI startup modeling computation on the human brain, announced $500 million in initial funding from Jeff Bezos, Lux Capital, and Google Ventures. Below them: Helion's $465 million Series G for fusion energy at a $15.5 billion post-money valuation led by Thrive Capital, longevity-medicine company NewLimit's $435 million Series C led by Founders Fund and co-founded by Coinbase's Brian Armstrong, AI music company Suno at $400 million, robotics startup Generalist AI at $400 million, market-intelligence platform AlphaSense at $350 million, and defense-tech company Mach Industries' $300 million Series C at a $1.8 billion valuation.

What Does It Change for Founders?

The week continues Q1's pattern with a difference worth noting: the money is spreading beyond frontier-AI model companies into the picks and shovels around them. Supabase (developer infrastructure), Ramp (finance automation), AlphaSense (market intelligence), and Impulse Space (space hardware) are all picks-and-shovels businesses in their domains — the kind of company whose value rises with usage of a bigger technology rather than by winning the model race itself. For founders outside the mega-AI club, that is the encouraging read: investors are paying growth valuations for companies that make AI-era work cheaper or faster without training their own frontier models. The caution remains the same as the quarter's: these are nine-figure rounds at multi-billion valuations, a market most startups never touch — the lesson to borrow is directional, the prices are not.

The takeaway for founders building around AI: the week of June 5 shows where the growth capital sits — infrastructure, automation, and hardware at the edges of the AI stack — and a $500 million round for a developer platform is the clearest signal yet that the picks-and-shovels thesis is the one investors are funding hardest.

Business News 7 publishes information, not employment or investment advice.

What Was Announced?

Per the May 8, 2026 roundup and the companies' statements: BILL announced cuts of up to 30 percent of headcount the same day as Cloudflare; PayPal outlined a reduction of about 20 percent of its roughly 23,800 employees — nearly 4,800 jobs — spread over two to three years, with CEO Enrique Lores saying the company would flatten its structure and accelerate AI adoption and automation; and Ticketmaster reportedly cut about 8 percent of its global workforce, roughly 350 employees across 25 countries. The stated rationales converge on the same three words: smaller teams, fewer layers, AI doing more of the work. Industry trackers counted more than 175,000 tech workers laid off across 2026 to date at the time of the wave.

What Does It Change for Founders and Job Seekers?

For founders, the wave validates a structure many startups already run — compressed teams with AI embedded in product and operations — and it lowers the cost of the argument with investors, who have spent 2026 asking portfolio companies to show AI-driven operating leverage before funding headcount growth. For employers hiring, the market has flipped in their favor at experienced levels: product, support, and engineering talent displaced from large companies is available at prices startups could not touch two years ago, and hiring processes that previously lost candidates to big-tech offers are closing. For workers and founders alike, the honest caveat is the one the reporting itself makes: research to date shows little evidence of broad AI-driven job disruption economy-wide — these announcements are corporate restructurings with AI as the stated rationale, not yet a measured labor-market shift. The founders reading the tea leaves correctly are treating the moment as both a hiring window and a positioning pressure: small, AI-leveraged teams are now the default benchmark against which every company's cost structure is compared.

The takeaway for owners and founders: a week of 14-to-30 percent cuts at major platforms resets talent supply and investor expectations at once — the winning move is hiring the displaced expertise while being able to honestly show the lean AI-era operating model investors now assume.

Business News 7 publishes information, not investment advice; funding figures are market data, not recommendations.

What Drove the Record?

Scale, not breadth. Four of the five largest venture rounds ever closed in the quarter — OpenAI at $122 billion, Anthropic at $30 billion, xAI at $20 billion, and Waymo at $16 billion — together accounting for roughly $188 billion, about 65 percent of the quarter's total, with ten additional companies raising billion-dollar-plus rounds. U.S. companies captured $250 billion, 83 percent of the global figure; China followed at $16.1 billion and the U.K. at $7.4 billion. The stage breakdown shows where the money actually went: late-stage funding hit $246.6 billion, up 205 percent year-over-year, while early-stage reached $41.3 billion, up 41 percent, and seed $12 billion, up 31 percent — with seed deal count actually falling about 30 percent.

What Does It Change for Founders?

The environment is two markets at once. For founders of AI companies and late-stage businesses with proven traction, the market is the most liquid in venture history — record sums chasing a defined set of themes, with valuations to match. For everyone else, the record headline overstates the weather at the bottom of the market: seed dollars rose modestly while seed deal count fell by nearly a third, meaning fewer checks concentrated in fewer companies. Practical readouts follow from that split — seed founders should assume competition for dollars remains sharp despite the aggregate record; later-stage AI-adjacent companies should price ambition accordingly; and founders outside AI should not read the $300 billion as evidence their own raise has gotten easier. The consistent Crunchbase-series caution applies to all: mega-rounds dominate the totals, so headline figures describe a market most startups do not inhabit.

The takeaway for founders: the Q1 2026 record is real but narrow — 80 percent AI, two-thirds in four companies' rounds — so the practical question for any raise is which of the two markets the company is actually in.

Business News 7 publishes information, not tax advice; 83(b) decisions belong with a tax professional.

Who Faces the 83(b) Decision?

The election applies to restricted stock — actual shares subject to vesting or forfeiture — not to stock options. The classic recipients are founders at incorporation (whose shares are worth almost nothing and who should file essentially without exception) and early employees granted restricted stock instead of options, common in the first hires before a formal option plan exists. Option holders face the decision only when exercising early: exercising unvested options makes the recipient an owner of restricted stock, and the 83(b) clock starts at exercise. Recipients of stock through repurchase-right founder agreements — including those with vesting imposed by investors — face it as well. For recipients of plain options who do not exercise, there is nothing to elect.

What Does Filing Actually Change?

Without the election, each vesting date is a taxable event: the recipient recognizes ordinary income equal to the shares' fair market value that day minus anything paid. For a company whose value rises, that means tax bills on paper wealth the recipient cannot spend — shares not yet sellable, taxed at ordinary rates as they vest, a problem that compounds into a genuine crisis when a company's value runs up before liquidity. With the election, the entire grant is taxed once at grant-date value — for early-stage companies, often pennies per share, producing a de minimis tax bill — and every subsequent gain is capital gain, at preferential rates, when the shares are eventually sold, with the holding period starting at grant. The trade is honest: the recipient pays tax on value that might evaporate, and cannot claim a loss for the amount included in income if the shares become worthless. For near-zero-value grants, that trade is trivially favorable; for grants with real value, it deserves actual computation.

How Is the Election Made?

Mechanics are strict. The election is a signed statement identifying the taxpayer, the property, the transfer date, the restrictions, and the value included in income — the IRS and several publishers provide template wording. It must be filed with the IRS within thirty days of the grant date: postmarked within the window, mailed to the service center, with proof of mailing kept forever. Copies go to the employer and are attached to the recipient's next tax return. The thirty-day deadline has no extensions and no late relief — courts have consistently upheld the deadline — which is why the standard practice is to prepare the election the same day the grant documents are signed. Companies onboarding recipients of restricted stock should hand each one the form and the instructions; recipients should not assume anyone else handled it.

What Happens Without One?

The failure mode is expensive and common enough to be a startup ritual: an early hire with restricted stock skips the election, the company's 409A value climbs through successful rounds, and each vesting tranche triggers ordinary income tax on appreciation the hire has not banked. At acquisition, the shares still pay off, but the tax paid along the way exceeded what capital treatment would have cost, sometimes by large multiples. The lesson: the 83(b) window is one of the few truly irreversible thirty-day deadlines in tax — identify whether a grant is restricted stock, decide quickly, and file with proof on day one.

Business News 7 publishes information, not legal advice; cap table corrections belong with startup counsel.

What Does Fully Diluted Actually Mean?

A fully diluted cap table assumes every right to shares is exercised and converted: issued shares, granted options (vested and unvested), the unallocated pool reserved under the equity plan, SAFEs and convertible notes at their conversion terms, and warrants. The distinction matters because investors calculate price per share and their ownership on fully diluted shares — a founder presenting only issued shares is presenting a smaller denominator, and the correction discovered at diligence reads as either sloppiness or concealment, both expensive. The table should also show the pro forma: what it looks like after the new money and after the SAFE stack converts, because that is the actual question on the table — what the founders will own after the deal everyone is negotiating.

What Are the Standard Diligence Failures?

Four patterns recur. Phantom promises: advisor or early-hire equity discussed verbally, sometimes even in emails, never granted — resurfacing as claims when the company is worth something. Stale pool math: a plan pool shown as available when grants against it were never papered, or vice versa. Missing 409A and grant records: options granted without a contemporaneous fair-market-value determination create tax exposure for the recipients and diligence work for everyone. Unreconciled registries: the spreadsheet, the transfer ledger, and the stock plan administrator's records disagree about who owns what. Each failure is fixable before diligence and reputation-damaging during it — the fix is a records audit against the source documents: board consents, grant agreements, the registry, plan documents.

How Should the Option Pool Be Presented?

The option pool deserves its own section because it is simultaneously real dilution and negotiation leverage. Investors typically require the pool be sized into the pre-money — meaning existing holders, usually founders, absorb the dilution of the pool before the new money arrives — and negotiate its size to cover the hiring plan through the next raise. The table should therefore show the pool both as-is and as-proposed, with the hiring plan that justifies the number, because "we need 15 percent" without a plan invites the investor to impose a smaller pool or a different structure. Founders should also understand the sequencing point: pool created pre-round dilutes founders; pool created post-round dilutes everyone including the investor, which is why investors prefer the former and model it.

What Format and Hygiene Win Diligence?

The winning presentation is boring: a current table from the company's equity-management platform or a clean spreadsheet reconciled to records, a security-by-security schedule (holder, amount, price, cap, vesting, documents referenced), the fully diluted count, and the pro forma with the contemplated round modeled at the discussed terms. Versioned files, a named owner for the records, and quarterly reconciliation turn diligence from archaeology into confirmation. The lesson: the cap table is the company's title deed — investors are not buying the story, they are buying the shares, and the table is where the shares live.

Business News 7 publishes information, not legal or financial advice; financing instruments deserve counsel review.

What Happens at the Priced Round?

When the company sells preferred stock at a fixed price per share, each SAFE converts by buying that round's preferred stock — or an equivalent series — at the SAFE price instead of the round price. The SAFE price comes from two mechanisms, and the investor gets whichever is cheaper. Valuation cap: the SAFE price equals the cap divided by the company's pre-money fully-diluted capitalization, so a $5 million cap on a company raising at a $25 million pre-money lets the SAFE buy at one-fifth of the round price. Discount: with no cap or where the round prices below the cap, the SAFE buys at the round price minus a stated discount, commonly 10-20 percent. A $500,000 SAFE with a $5 million cap converting in a round priced at $25 million pre-money buys $2.5 million worth of round shares at the SAFE's $500,000 cost — a 5x paper multiple at signing, before any later outcomes.

What Do the Different Flavors Change?

SAFEs come in cap-only, discount-only, cap-and-discount (MFN between them), pre-money and post-money variants, and the post-money variant Y Combinator standardized in 2018 changes the founder math enough to matter. A post-money SAFE converts on the company's capitalization after the round is added but before the new money — meaning the SAFE percentage is fixed and knowable at signing, and successive SAFEs dilute only the founders, not each other. Pre-money SAFEs, the older form, share dilution among all SAFE holders and compute against the capitalization including converted SAFEs. Founders stacking multiple SAFEs should model both: the same headline cap produces different founder ownership depending on which flavor sits in the stack, and "post-money" is not a better default — it is a different one.

What About Pro-Rata Rights, Liquidity, and Failure?

Conversion brings details worth knowing. Most SAFEs carry pro-rata rights letting the investor participate in the converted round alongside new money, which preserves ownership through the dilution their own conversion causes. SAFEs have no maturity date and no interest — they sit indefinitely until a priced round, a sale, or a dissolution. In an acquisition, SAFEs typically convert to common or take their money back at a negotiated multiple, whichever the documents provide; in a dissolution, SAFEs stand behind almost everything — their payout is what remains after creditors and preferred preferences, usually nothing. MFN clauses let early SAFEs adopt better terms offered to later SAFE investors before the priced round, which is why SAFE stacks tend to converge on the best terms signed.

What Should Founders and Investors Model?

Founders should run the stack before any priced round: each SAFE's conversion shares, the dilution to existing holders, and the effective price the new investor's money pays after SAFE dilution — because sophisticated lead investors will do exactly that arithmetic and price it in. Investors should model ownership as of conversion, not as of signing, remembering that rounds and additional SAFEs signed later can shift pre-money flavor percentages. Both sides should treat the cap as the real negotiation: everything else in a SAFE is boilerplate, but the cap (and post- versus pre-money) is where the company is bought. The lesson: SAFEs defer the negotiation, not the economics — the price was set the day the cap was.

Business News 7 publishes information, not financial or tax advice; exercise decisions carry tax consequences and belong with a professional.

Why Do Options Go Underwater?

A stock option's value is the spread between the strike price — fixed at grant, set at the then-current 409A fair market value — and the share's actual worth. When the preferred price falls from, say, $10 to $5, options granted at strikes near $10 are underwater: exercising buys shares for more than they are worth. The practical consequences are sequential. New grants reprice automatically, because the company's independent 409A valuation resets, so new hires receive lower strikes than employees who joined earlier at the peak — a morale inversion familiar to anyone who lived through 2022-2023 or 2008. And the company's retention economics break: equity that stops being a motivating asset pushes exactly the people it was meant to retain toward the door.

What Do Liquidation Preferences Have to Do With It?

Down rounds frequently bring new money in ahead of the common stock: liquidation preferences — the investors' right to take their money (sometimes a multiple, sometimes with participation) off the top in a sale — stack. A company that raised $100 million across rounds and sells for $110 million may return little to nothing to common shareholders after preferences, even though headline valuations once implied riches. This is the arithmetic employees should run before weighing any offer or retention grant: the exit math is waterfall math, and preferences from a down round sit in the waterfall ahead of employee shares. Companies are not required to disclose the full preference stack, but candidates can and should ask for it — the cap table's total preference is a fair question in any negotiation.

What Are Reprisings and Refresh Grants?

Boards respond to underwater equity in two ways. A repricing amends existing option strikes to the current lower 409A value — often with vesting resets or conditions attached, and under accounting rules requiring expense re-measurement that boards dislike, which is why repricings cluster in severe downturns. More common are refresh grants: leaving old options alone and granting new ones at the low strike, layered on additional vesting. Employees comparing the two should note the asymmetry: a repricing preserves your existing schedule at a fair strike, while a refresh restarts a vesting clock — valuable, but not equivalent. Either way, the moment to ask is when the company needs retention most, which is exactly after a down round.

What Should an Employee Actually Do?

A short playbook beats panic. Get the facts: current 409A price, total preference stack, and any refresh or repricing policy. Model exits at several prices rather than the last preferred round's fantasy. Watch the tax mechanics — incentive stock options exercised while underwater still trigger alternative minimum tax calculations in some situations, and post-grant price declines interact with 83(b) timing and ordinary income treatment in ways a tax adviser should price. Consider the recovery path: companies that raise down and survive often deliver real value to refreshed equity, and employees who leave at the bottom forfeit unvested recovery upside for a lateral move. The lesson: a down round resets the bet, it does not end it — and employees who know the difference between strikes, 409A values, and waterfall proceeds negotiate the right half of the problem.