Fast Five (the Fed decision, BA, BE, different take on MSFT, SOFI)
Fast Five is a weekly snapshot, not a deep dive. We pick five of the most important earnings releases from the week, pull out the numbers and moments that actually matter, and lay them out simply and fast. Companies are listed alphabetically, and that is intentional. What moves the needle for one investor may be completely irrelevant to another. If a name catches your eye, you will already know where to start.
For each company I will also provide my take on what to do with the stock. I have chosen to summarize the results into the following 3 categories:
High Conviction: strong results, clear narrative, worth digging deeper into
Worth Watching: interesting but not conclusive, something to monitor
Proceed with Caution: results or signals that warrant skepticism before going deeper
Without further ado, let’s bust right into it, but first:
The Fed: Standing Still But Progress Being Made
Rates were held unchanged at 3.5% to 3.75%. No surprise on the decision but surprise on the dissent count. Three regional Fed presidents (Logan, Hammack, and Kashkari) voted for a 25 basis point rate hike. That is a committee that is genuinely split down the middle, with nine officials looking for no change or a cut and nine others looking for one or more hikes in 2026. Warsh held rates, but he is presiding over a divided institution. The unanimous statement from June was a one-meeting phenomenon, not a trend.
The most important thing Warsh said was not about rates, it was about the bond market. Warsh repeatedly welcomed the increase in bond yields since the June 17 FOMC meeting, driven by real yields, suggesting he sees the market itself doing the tightening work even with no change in the Fed funds rate. This is Warsh’s operating philosophy made explicit. He removed forward guidance so markets would react to data rather than Fed signalling. The bond market repriced significantly since June and in Warsh’s view, that repricing is tightening financial conditions, which means the Fed can afford to wait.
Broad financial conditions, however, remain easy. Equity markets have moved sideways, credit spreads remain tight, the housing market and low-income households face restrictive conditions while AI-led business investment faces few credit constraints. That unevenness is the K-shaped economy playing out in monetary policy. Rates are high enough to hurt some people but they are not high enough to slow the parts of the economy that are driving inflation.
Warsh was emphatic about price stability but revealed almost nothing about his reaction function, namely what he would specifically need to see to either hike or hold in September. He rejected pure data dependency, saying “data need to be broad” and that there was “a problem with being dependent.” He hinted at looking through tariff and supply shock inflation to focus on underlying trends. Warsh is the most opaque Fed chair in modern memory and he wants markets guessing.
AXA Research’s baseline is that the Fed remains on hold in September, provided inflation expectations stay anchored and underlying inflation shows signs of decelerating over the next couple of CPI and PCE reports. A rate hike would require the combination of a tightening labour market, a renewed increase in oil prices, persistent tariff-related inflation, or strong AI-linked price pressures. AXA sees CPI averaging 3.6% in 2026 and 2.4% in 2027, slow deceleration, but deceleration nonetheless. The next key dates are the July and August CPI prints and the June and July PCE releases. Those four reports will tell us whether Warsh’s patience is vindicated or whether the three dissenters get their hike in September.
Rating: This Is Not A Neutral Fed. Rates were held but three officials voted to hike, the bond market is doing the tightening work, and Warsh has given no clarity on what moves him in either direction. The base case is hold in September. The environment is tightening it’s just the mechanism that has changed.
Boeing (BA)
Boeing reported Q2 revenue of $24.56 billion, up 8% year over year and ahead of estimates. The headline EPS looked ugly, an adjusted loss of $0.76 against a consensus of $0.30, but buried inside that miss was a $280 million charge on the Air Force One replacement program, a fixed-price government contract that has been a financial sinkhole for years. Strip it out and this is a company that is quietly, measurably getting better. The stock rose more than 4% anyway.
The number that matters is free cash flow. Boeing generated $631 million in positive free cash flow, swinging from a $200 million burn in the same quarter last year and obliterating the $177 million burn analysts had expected. For a company that hemorrhaged cash through strikes, safety scandals, production freezes, and quality crises, this is not a footnote. It is a turning point. Consolidated debt fell from $47.2 billion to $45.9 billion in a single quarter. Boeing is now paying down debt so that is a different company than the one that existed twelve months ago.
The operational recovery underneath the financials is real. 171 commercial deliveries, up 14% year over year, was Boeing’s highest quarterly total since 2018, bringing first-half deliveries to 314, the strongest first half in eight years. The 737 program transitioned to 47 aircraft per month. Certification testing is complete on both the 737-7 and 737-10. A fourth 737 assembly line activated in July. The production machine is turning back on, one line at a time.
The risks are still real and worth naming. The 777X received FAA approval to begin certification flight testing but first delivery remains 2027, following a $4.9 billion charge recorded last year. The Air Force One program is still bleeding. And $45.9 billion of debt does not disappear in a quarter. CEO Kelly Ortberg said it plainly: “While two quarters don’t make a year, if we work together and stay focused on safety, quality and on-time performance, we’ll set ourselves up for a big second half.”
The total backlog hit a record $715 billion, including more than 6,200 commercial aircraft. Full-year free cash flow guidance stands at $1 billion to $3 billion. Nine years of production already sold. The demand was never the problem. The execution was. Ortberg is fixing the execution.
Rating: Worth Watching, trending toward High Conviction. Boeing is not fixed. But it is fixing. Free cash flow turned positive, deliveries are accelerating, debt is declining, and the backlog is the largest in company history. If the full-year free cash flow guidance holds, this stock has a long runway from here.
Bloom Energy (BE)
Bloom Energy reported Q2 revenue of $1.065 billion, up 165.5% year over year, its first quarter above $1 billion in company history. Non-GAAP EPS of $0.78 beat the $0.40 consensus by 95%. One of the best beats this quarter. After a brutal July that saw shares fall 40% from their highs on a short-seller attack, the numbers answered the most important question: the business is real and it is accelerating.
The context matters here. Bloom’s stock had rallied 247% in the first half of 2026 on surging AI data center demand. The collapse began July 8 when short-seller Hunterbrook Media published “Bloom’s Big Lie,” alleging the company sourced scandium oxide, a critical material in its fuel cells, from Chinese suppliers through intermediary countries, despite CEO K.R. Sridhar’s repeated claims of no China supply chain.
Product revenue surged 215% to $935 million. Non-GAAP gross margin expanded to 34.3%, up 604 basis points year over year. Real volume, real margins, real operating leverage. Brookfield simultaneously announced an expansion of its financing framework for Bloom’s AI power projects from $5 billion to $25 billion. When one of the world’s largest alternative asset managers quintuples its capital commitment in nine months, that is a credibility signal no short-seller report can easily dismiss.
Full-year guidance was raised to $3.9 to $4.2 billion in revenue, roughly 100% growth at the midpoint, with non-GAAP operating income of $800 to $900 million. Third consecutive quarter of raising the bar. But two legitimate risks remain.
The scandium sourcing question is not fully closed despite management’s SEC denial.
The backlog disclosure gap, a “$20 billion contracted backlog” cited publicly versus $492 million in audited remaining performance obligations in SEC filings, deserves a clear explanation, not a dismissal.
Bloom sits at the intersection of two unstoppable forces: AI’s insatiable power demand and the grid’s inability to meet it fast enough. Hyperscalers cannot wait years for new grid connections. Bloom bypasses the grid entirely. Management claims Bloom has become the standard for on-site power in the AI data center market within a year. The demand is real. The transparency questions are the ones that could still derail it.
Rating: Worth Watching with High Conviction potential. The numbers are extraordinary and Brookfield’s commitment is meaningful third-party validation. But the sourcing controversy and backlog disclosure gap are real risks. Watch supply chain transparency and H2 backlog conversion before sizing up fully.
Microsoft (MSFT)
Microsoft reported Q4 FY2026 revenue of $90.01 billion, up 17.8% year over year and ahead of the $87.63 billion consensus. Non-GAAP EPS of $4.74 beat the $4.24 estimate by nearly 12%. Azure grew 43% in the quarter and crossed $100 billion in annual revenue for the first time. Commercial remaining performance obligations — essentially contracted future revenue — surged 84% to $678 billion. Full fiscal year revenue surpassed $331 billion. On paper, one of the strongest quarters in Microsoft’s history. But the numbers are not the story, the accounting is.
Starting in fiscal 2027, Microsoft is extending the estimated useful life of its data centres and office buildings from 15 years to 25 years. That is not an operational decision, that is an accounting decision. The mechanics are straightforward: longer useful life means lower annual depreciation, which means higher reported earnings. But the change does not stop there. That shift in useful-life assumption also changes how certain leases are classified under accounting rules. Data centre leases that were previously classified as finance leases (which show up in reported capital expenditure) will now be treated as operating leases, which do not. The result: Microsoft’s headline 2026 capex guidance falls from roughly $190 billion to around $175 billion, even though the underlying spending plans have not changed.
This matters because the numbers underneath the accounting change are staggering. Capital expenditures and finance leases for Q4 reached $41 billion, up 69% year over year. Over the last four quarters, Microsoft has spent approximately $97 billion on infrastructure and equipment to generate roughly $37 billion of annual recurring AI services revenue. The spend is accelerating. FY2027 capex guidance has now been disclosed at $255 to $260 billion, a 35% increase over 2026 levels and significantly higher than prior analyst expectations. And management noted that around $25 billion of the increase is driven by higher component costs (GPUs, memory, and other materials) rather than simply buying more capacity. In other words, the same number of data centres costs $25 billion more than it did a year ago.
Here is the question that the accounting change cannot answer. The useful-life extension assumes Microsoft’s data centres will be productive assets for 25 years, yet the GPU compute hardware inside those data centres has an effective useful life of around six years. The economics and the accounting are running on different clocks. Investors who look only at the headline EPS beat will not see it but Investors who read the footnotes will.
Three simple questions should guide how you think about Microsoft’s AI bet.
why change the useful-life assumption now, at the exact moment that capex is becoming politically and analytically uncomfortable?
does $97 billion of infrastructure spending to generate $37 billion of AI revenue represent an acceptable return trajectory, and over what time horizon does that math improve?
if FY2027 capex rises to $255 to $260 billion, what does the free cash flow statement look like when a significant portion of that spending migrates off the capital expenditure line and into operating leases
Rating: Worth Watching. The Azure business is genuinely exceptional and the commercial backlog of $678 billion provides real forward visibility. But $255 to $260 billion of annual capital commitment, a useful-life accounting change that flatters both earnings and reported capex simultaneously, and infrastructure spending that is outpacing AI revenue by more than two to one. The business is extraordinary but the capital allocation requires scrutiny that the headline beat does not invite.
SoFi Technologies (SOFI)
SoFi reported Q2 GAAP net revenue of $1.22 billion, up 43% year over year, with net income of $156.6 million and diluted EPS of $0.12, beating the $0.11 consensus. Loan originations hit a record $14.8 billion and net income surged 61% year over year. Full-year adjusted net revenue guidance was raised to $4.75 to $4.85 billion, above the prior $4.7 billion analyst consensus. Beat, raise, record originations, eleventh consecutive quarter of GAAP profitability.
The lending segment posted adjusted net revenue of $711.7 million, up 59% year over year. The financial services segment generated $466.3 million, up 29%. Two of three business segments are firing. The one that is not: the technology platform segment posted net revenue of $84.5 million, a 23% decline, reflecting the loss of a large client that departed the platform before the end of 2025. That client departure has been a persistent overhang on the stock and the segment’s recovery trajectory will matter for how the market re-rates SoFi from here.
SoFi added a record 2.2 million products in Q2, a 42% increase year over year, bringing total members to 15.8 million, up 35% from a year ago. The number that CEO Anthony Noto called a “major milestone”: SoFi is now adding twice as many products as members each quarter. That is the everything-app thesis working in practice. Members are not joining SoFi for one product and leaving. They are joining for one and staying for two, three, or four. 51% of new products opened in Q2 were by existing members, up from 43% last quarter and 35% a year ago.
The interest rate environment is both SoFi’s tailwind and its risk. Demand from borrowers refinancing high-cost debt, funding education, and seeking home financing has driven record originations for two consecutive quarters. Credit quality is the other variable. CEO Anthony Noto said “spending remains strong, demand remains strong, and credit performance continues to meet or exceed our expectations.” The question is whether it remains true if unemployment ticks up or consumer balance sheets deteriorate.
Rating: High Conviction. Eleven consecutive quarters of GAAP profitability, record loan originations, 40% revenue growth, rising cross-sell ratios, and a full-year guidance raise. The tech platform headwind and rate sensitivity are real risks. The valuation at these levels prices in a lot of bad news that the actual results do not support. The gap between the business and the stock price is where the opportunity lives.
Disclaimer: The views and opinions expressed above are current as of the date of this document and are subject to change without notice. Materials referenced above are provided for educational purposes only. Nothing above constitutes investment advice, a recommendation or an offer to sell, or a solicitation of an offer to buy, any securities or investment products. Always conduct your own due diligence and consult a qualified financial professional before making investment decisions.