Tyler Jacobsma is the founder of Flowframe.xyz, which provides in-depth content and tools for prediction market traders.
Wednesday afternoon is going to be absurd. Four companies worth a combined $10+ trillion in market cap will release their quarterly results within about an hour of each other, and every single one of them is a data point on whether the AI spending spree is translating into actual earnings growth. This is a stress test for the entire market narrative around artificial intelligence.
Here's Polymarket's "will they beat EPS?" scoreboard:
| Company | EPS Estimate | Polymarket: Beats? |
|---|---|---|
| Amazon (AMZN) | $1.65 | 96% YES |
| Microsoft (MSFT) | $4.05 | 94% YES |
| Meta (META) | $6.62 | 93% YES |
| Google (GOOGL) | $2.66 | 96% YES |
Every single one is above 90%. Polymarket traders are very confident that all four companies will beat their EPS estimates. At first glance, that feels too high, right? 93-96% is the kind of number you see on near-locks, but earnings can be volatile, and analyst estimates are not always accurate. (As we saw last week with Tesla, which had only a 20% chance to beat on Polymarket, but it ended up clearing the hurdle anyway)
You might be asking, "Why would beating earnings be a good 90%+ bet?"
Because beating EPS estimates is actually the norm for these companies, not the exception. Over the last 20 quarters, the S&P 500 "beat rate" on EPS has averaged about 77%, and mega-cap tech companies beat at an even higher rate because analysts tend to sandbag their estimates conservatively.
Amazon has beaten EPS in 18 of the last 20 quarters, and Microsoft has a similar streak. Meta and Google rarely miss on the bottom line, even when they disappoint in certain areas.
When looking at earnings on Polymarket, it’s not usually about “did this company have a good quarter” because there are a lot of variables that go into that. It's asking a more specific question: "Will reported EPS come in above $1.65?" Given that these large tech companies have a high beat rate and the way analysts come up with their estimates, 96% is aggressive but not crazy.
The Stock Doesn’t Care About the Beat. It Cares About These Numbers
Beating EPS at 93-96% probability means the prediction market has already priced in the beat, and so has the broader finance world. If you buy YES at 96 cents, you're risking 96 cents to make 4 cents. That's not a great risk-reward trade even if you're right.
Last quarter, Amazon beat EPS estimates, and the stock dropped 6% the next day because investors hated the $200 billion capex guidance, while Meta beat and fell on free cash flow concerns. Beating their analyst estimate but getting punished for the outrageous spending outlook has become a pattern of this AI cycle. The market barely rewards revenue because it’s expected, but it punishes spending.
Here's what actually will move each stock on Wednesday night, regardless of whether the EPS beat hits:
Amazon ($1.65 EPS est.) — AWS growth rate is the only number that matters. It grew 24% in Q4 2025, the fastest in 13 quarters. Morgan Stanley is modeling 29% for Q1 and 31% for the full year. If AWS comes in below 25%, the stock sells off no matter what headline EPS says. Amazon also guided capex at $200 billion for 2026, which could push free cash flow negative by as much as $28 billion, according to Bank of America. CEO Andy Jassy will need to convince investors that spending into a hole produces returns on the other side.
Microsoft ($4.05 EPS est.) — Azure constant-currency growth, guided at 37-38%. Azure grew 39% last quarter, so the bar is high. Microsoft 365 Copilot at $30/user/month is the company's biggest near-term AI monetization bet, and any acceleration in enterprise seat count moves the stock more than the EPS number. Microsoft is down 17% year-to-date, the worst of the group, because investors aren't sure the AI spending (a record $37.5 billion last quarter alone, up 66% YoY) is translating fast enough.
Meta ($6.62 EPS est.) — Ad revenue powered by Andromeda, Meta's AI ad-targeting engine. Wall Street expects $55.5 billion in Q1 revenue with 32% growth. The cleaner story here is that ads still work and they're funding the AI build-out, but capex jumped to $115-$135 billion for 2026, and Barclays now projects negative free cash flow in 2027 and 2028. Meta has zero Sell ratings across 42 analysts. When consensus is unanimous, any disappointment gets amplified.
Google ($2.66 EPS est.) — Google Cloud growth is the headline. It surged 48% in Q4 2025, pushing the segment to a $70 billion annual run rate, with a $243 billion backlog that gives Alphabet unusual revenue visibility. Search growth around 17-18% is the secondary metric. Morgan Stanley analyst Brian Nowak flagged that the market "is not pricing in any disruption risk" for Google — meaning if AI Overviews start cannibalizing ad clicks instead of boosting them, the repricing is swift.
The takeaway
The Polymarket odds at 93-96% are telling you something simple: these companies will almost certainly report a number higher than the estimate. That's just how mega-cap earnings usually work.
The more interesting thing to watch is whether the market rewards the beat or punishes astronomically high spending guidance and what the new narrative around AI capex will be come Thursday morning.
