Executive compensation changed dramatically in early 2026. Meta and Tesla are betting billions that all-or-nothing pay packages are the only structure that can compete in the AI talent market.
TL;DR
- Meta issued stock options to top executives for the first time since its 2012 IPO, with strike prices requiring the company to reach a market cap of $8–9 trillion for full payout.
- Tesla shareholders approved a pay plan that requires $7.5 trillion in new value creation, plus 20 million vehicles, 1 million Optimus robots, and 1 million robotaxis.
- Mark Zuckerberg has personally hand-delivered soup to AI researchers he wants to recruit, while Meta reportedly offered one hire a $1.5 billion package.
- Companies that use Economic Profit in executive incentives beat the S&P 500 by 7% annually and outperform peers on TSR by 4.7 percentage points.
- Proxy advisory firms like ISS and Glass Lewis face growing criticism from Congress and the White House for enforcing standards that penalize bold compensation plans.
- The fix for broken executive pay: indexed options that filter out market luck, uncapped profit-sharing (like Nucor’s model), and single-plan architectures that force clarity.
Written by David Berkowitz, founder of VAP Wealth Advisors and creator of Berk on Value. David manages the ValueAligned Portfolio (VAP) model portfolio for long-term investors. His research focuses on owner-oriented capital allocation, economic profit analysis, and identifying compounding businesses that create real shareholder value. For more, visit vapwealthadvisors.com.
Meta just told six executives: make us worth $9 trillion or get nothing
In March 2026, Meta Platforms did something it hadn’t done since its IPO fourteen years earlier. It granted stock options to senior executives. Not restricted stock. Not time-vested RSUs. Options with strike prices that start 88% above the current stock price and climb to a 6x multiple.
Six people received the grants: CTO Andrew Bosworth, CFO Susan Li, CPO Chris Cox, COO Javier Olivan, CLO C.J. Mahoney, and President Dina Powell McCormick. Mark Zuckerberg was excluded. His 13%+ ownership stake already ties him to every dollar of shareholder value the company creates or destroys.
The lowest tranche kicks in at $1,116.08 per share. The highest demand is $3,727.12. For that top tranche to pay anything, Meta would need to reach a market capitalization of $8–9 trillion—a 500%+ increase from its current $1.5 trillion baseline. If the stock doesn’t hit the targets before February 2028, the options expire worthless.
The board is telling its leadership team directly: $115–135 billion in 2026 capex must produce returns that justify a market cap larger than the combined current GDP of Japan and Germany.
The options rationale
Soup wars. OpenAI’s chief research officer, Mark Chen, revealed that Zuckerberg personally hand-cooked and hand-delivered soup to researchers he wanted to recruit. Meta went after half of Chen’s direct reports. The company reportedly offered one AI researcher, Andrew Tulloch, a package worth as much as $1.5 billion over six years.
Standard RSUs can’t compete with that kind of arms race. RSUs reward tenure. Options reward value creation above a specific price target. When competitors hand out nine-figure packages, the only credible counter is a structure that offers exponentially more—but only if the stock price proves the bet was right.
Tesla’s answer: $8.5 trillion or bust
Tesla’s board faced a different version of the same problem. Elon Musk wanted 25% voting control to protect the company’s AGI and robotics ambitions from activist interference. The board’s response was the 2025 CEO Performance Award, approved by over 75% of shareholders at the November 2025 annual meeting.
The structure: 423 million shares of restricted stock split across 12 tranches. An “Offset Amount” equal to the grant-date stock price ($334.09) gets subtracted from every share at settlement. That transforms a full-value grant into a growth-only instrument. If Tesla’s stock doesn’t appreciate, Musk gets zero.
Full payout requires Tesla to reach $8.5 trillion in market cap—roughly $7.5 trillion in new value from its current ~$1 trillion baseline. The operational milestones are equally demanding: 20 million vehicle deliveries, 1 million Optimus humanoid robots delivered, 1 million robotaxis in commercial operation, and 10 million active FSD subscriptions.
| Dimension | Meta 2026 Plan | Tesla 2025 Award |
| Vehicle type | High-hurdle stock options | Restricted stock with Offset Amount |
| Recipients | 6 C-suite executives (no CEO) | Solo founder-CEO |
| Performance hurdles | Pure price/market cap targets | Dual: market cap + operational |
| Max market cap target | $8–9 trillion | $8.5 trillion |
| Vesting window | Price targets by Feb 2028; time-vest through Aug 2030 | 7.5 and 10 years if milestones are met |
Two completely different structures. One shared philosophy: you capture nothing until shareholders capture trillions.
The AI talent economy is repricing every job in the C-suite
These plans didn’t emerge from boardroom theory. They’re a direct response to a labor market that has gone haywire. OpenAI’s average stock-based compensation hit $1.5 million per employee in 2025—the highest of any tech startup in history. Distinguished AI researchers command $1–$5 million in total annual compensation, with four-year packages exceeding $100 million at the top end.
| Role | Typical total comp range | Notes |
| Machine Learning Engineer | $180K – $350K+ | Baseline AI competency |
| Generative AI / LLM Engineer | $400K – $900K+ | 25–45% premium over non-AI tech |
| Distinguished AI Researcher | $1M – $5M+ | Can reach $100M+ over 4 years |
| Head / VP of AI | $700K – $2M+ | Leadership of AI pivots |
The gap between standard tech and AI tech compensation is now a 25–45% strategic premium ( Rise AI Salary Report 2026). When individual researchers earn more than most CEOs, executive pay must scale upward or the leadership team fractures. A board that pays its C-suite below the going rate for the people they manage will lose both.
The metric that separates real value from accounting fiction
Paying executives billions only makes sense if you can measure whether they’re creating or destroying value. GAAP doesn’t do this. Revenue growth can mask capital destruction. EPS inflates through buybacks funded by debt. The metric that separates real value creation from accounting distortion is Economic Profit: operating profit after tax, minus a charge for all capital employed.
A Fortuna Advisors study published in the Journal of Applied Corporate Finance examined 32 public companies that use EP in their executive compensation. The results: EP companies outperformed peers on annualized total shareholder return by 4.7 percentage points and beat the S&P 500 by 7.0% annually. Their EBITDA margins ran 3% higher than peers.
The reason is behavioral. When managers face a capital charge, they stop treating capital as free. Every dollar of investment must earn more than its cost. That single constraint kills the two most common forms of value destruction: empire building during booms and clinging to bad capital during busts.
Economic Profit v1.0 had a flaw. Eco nomic Profit v2.0 fixes it.
Traditional EP penalizes new investment because depreciation front-loads costs. A plant bought yesterday looks expensive on an EP scorecard. A plant bought a decade ago looks cheap because the asset base has depreciated. This creates a perverse incentive to sweat old assets rather than invest in new ones.
EP 2.0—the Fortuna model—corrects this with two adjustments. First, it uses undepreciated assets, meaning the original cost, so every dollar of capital faces the same charge regardless of age. Second, it capitalizes R&D and brand spending rather than expensing them immediately. GAAP treats a $10 billion AI research budget as a current-period cost. EP 2.0 treats it as a long-term asset and amortizes it, which stops penalizing companies that invest in their future.
For moonshot compensation plans, EP answers the question that the stock price hurdle alone cannot: whether growth is funded by genuine returns or by burning capital on projects that never pay off.
The governance machine that kills bold pay plans before they start
Public company boards don’t operate in a vacuum. They answer to proxy advisory firms—primarily ISS and Glass Lewis—that wield outsized influence over shareholder votes. These firms have a documented bias toward cost control and standardization over alignment and motivation.
A Congressional Research Service report found that ISS and Glass Lewis exert influence that “distorts governance decisions and undermines shareholder trust.” Companies change their compensation design specifically to avoid negative voting recommendations. Over a third of equity plans include share counts within 1% of ISS approval limits—meaning the advisory firm’s formula, not the board’s judgment, determines the size of the grant.
This creates what compensation researchers call the Lake Wobegon Effect: every board benchmarks to the 50th–75th percentile of peers, no board considers its executives below average, and the entire system ratchets upward in cost while converging on mediocrity in design. About 90% of major U.S. companies set pay targets at or above the peer median.
The resulting damage runs through seven common practices: rewarding revenue growth that buys sales at the expense of profit; layering multiple overlapping three-year plans that create zero priorities; deriving targets from annual budgets, which rewards sandbagging; and capping bonuses, which caps performance.
Private equity-backed firms avoid this trap entirely. Research from Boston University Law shows PE owner-directors use uncapped profit-sharing, large upfront equity grants, and subjective discretion to reward value creation without fear of a Wall Street Journal headline. That is a structural governance advantage public companies can’t easily replicate.
But they have to try. The Tornetta v. Musk ruling in Delaware showed what happens when moonshot plans come from a “cooperative venture” with a compromised board. The entire fairness standard demands an adversarial process, fully informed shareholder votes, and genuine independence on the compensation committee. Plans built on those foundations withstand legal scrutiny. Boards that negotiate with themselves do not.
Five moves smart boards are making right now
The most direct structural change is replacing budget-based targets with profit sharing. Nucor has run this model for decades: workers can double or triple base pay when the company is profitable, with no cap. The bonus pool draws from profits above a capital threshold, not from a negotiated budget. Managers can’t game what they can’t negotiate.
Indexed options address a separate design problem. Standard options pay executives when the whole market rises. Indexed options adjust the strike price against a peer group or sector benchmark, so the executive earns for outperformance—not for riding a bull market. Every dollar of option value traces back to decisions the executive actually made, cutting the “managerial discount” embedded in conventional grants.
The cap on incentive pay is the most underappreciated design flaw in corporate compensation. A cap on the bonus functions as a cap on the behavior. Boards worried about runaway payouts have better tools: deferral and clawback. Deferring a portion of above-target bonuses into restricted accounts that vest over three years protects shareholders without telling executives that a dollar of profit beyond the cap is worth nothing to them. If the results reverse, the money comes back.
Market-derived targets remove the sandbagging problem. Internal budget negotiations invite conservative projections. Targets set against improvement rates implied by the current stock price, or against peer-relative performance thresholds, eliminate the incentive to lowball. If the market prices in 15% earnings growth, that becomes the baseline. Results above it earn a bonus; results below it mean the executive underdelivered against what investors already paid for.
Simplicity is the final constraint. Three overlapping incentive plans with different metrics, timelines, and peer groups create confusion, not alignment. A single plan built around one long-term value metric—applied consistently—removes ambiguity about what the company actually rewards. Nucor’s model proves that simple designs and strong motivation reinforce each other.
What this means for your portfolio
When a board grants executives options at a 6x strike price, it is making a public statement about where it thinks the company is headed. That statement is worth reading carefully.
Meta’s board is telling shareholders it believes AI infrastructure spending will produce returns large enough to justify an $8–9 trillion valuation—and that the six leaders running the company are worth billions if they deliver. Tesla’s board is saying Musk can build a fleet of autonomous vehicles and humanoid robots worth $7.5 trillion more than today. Both bets could fail. The incentive structures guarantee that if they do, executives collect nothing.
For long-term investors, the question is whether these plans align executive behavior with shareholder outcomes. The data says they can—if the right metric sits underneath. Economic Profit forces capital discipline. Indexed options filter out market luck. Uncapped, simplified plans keep managers focused on the work that actually creates value. The boards that get this right will attract the talent the AI economy demands. The ones that hide behind proxy-advisor templates will watch that talent walk out the door carrying a $1.5 billion offer from someone who doesn’t.
