The alpha Brief: Insights on Quant Investing and the alpha Brief Portfolio (aBP)

A strong investment thesis should become stronger when new evidence arrives. This week, an AI-connectivity company translated product momentum into earnings leverage, a defense contractor expanded backlog visibility through a major missile award, and a gold miner combined lower costs with record profitability and aggressive buybacks. The common thread wasn't the share-price reaction. It was the growing body of evidence behind each long-term investment case.

I downgrading AMD from a STRONG BUY to a HOLD. The Q2 evidence supports the product thesis, but only 0.4% upside from the $549.88 intraday price leaves too little margin of safety before the MI450 and Helios ramp is proven. AMD reported Q2 2026 revenue of $11.536 Bn, up 50% y/y, and Non-GAAP EPS of $1.66. Revenue beat the $11.31 Bn consensus by about 2%, while EPS beat the $1.61 consensus by 3%.

I am downgrading Corning from STRONG BUY to a SELL. The business is executing, but the stock still prices in too much of the 2028 to 2030 opportunity, leaving about 24% downside to my base target. Corning delivered Q2 core sales of $4.738Bn and core EPS of $0.78, above consensus of $4.63Bn and $0.755. Core operating margin expanded to 20.9%.

Q2 revenue increased 62% year over year to $4.70 Bn and adjusted EPS increased 83% to $2.54, beating consensus by about $300 Mn and $0.23, respectively.

I maintain a Buy rating on MCHP since the operating recovery is stronger, but approximately $5.1Bn of net debt, preferred-stock obligations and dilution keep this from being a risk-free rerating story.

Fiscal Q4 revenue rose 30% year over year to $6.72 Bn and adjusted EPS reached $1.82, beating consensus by about $57 Mn and $0.13, respectively.

Not all corporate announcements deserve the same weight. This week, one company extended a production platform, another converted proven technology into international demand, and a third introduced an ambitious AI architecture that still needs real-world deployment. The question isn't which story sounds biggest. It's how much proof already exists and how much still depends on execution.

I am downgrading Sandisk from a STRONG BUY to a HOLD. The business is stronger, but I think that 9% base-case upside at $1,605 is not enough for a BUY given NAND pricing and peak-margin risk. Sandisk reported Q4 FY2026 revenue of $8.97Bn and Non-GAAP EPS of $39.25, both above consensus, while non-GAAP gross margin reached 84.6%.

The AI-optics thesis strengthened: both Components and Systems grew more than 100% year over year, initial 1.6T shipments began, and fiscal Q1 guidance moved to $1.225-$1.275 Bn of revenue.

This was a useful week because the most important developments did not all arrive with a rising share price. Yes price tells us how the market reacted over five sessions but the company development tells us whether the underlying thesis became stronger, weaker, or simply more complicated.

I downgrade POET Technologies (POET) from Strong Buy to Hold. The balance sheet is much stronger, but the investment case still depends on qualification, production and revenue conversion that the latest quarter has not yet proved.

I am downgrading Ase Technology from a STRONG BUY to a HOLD. It's earnings execution is better than I expected, but negative free cash flow, higher net debt, and a near 30x FWD P/E leave little margin of safety.Q2 revenue rose 26.7% y/y to NT$191.1Bn and diluted EPS increased 171% to NT$4.61, or $0.292 per ADS. ATM revenue grew 36.3% and its operating margin reached 15.7%.

I downgrade HPE from Buy to Hold. The company is executing better than my old thesis, but the current price of $60.76 leaves almost no base-case upside. 3Q FY 2026 revenue reached $12.2Bn, up 34% Y/Y, while non-GAAP EPS of $1.11 beat the $0.93 consensus by $0.18.

I am downgrading Silicon Motion from a STRONG BUY to HOLD. Execution improved, but the prior valuation discount has disappeared and first-half free cash flow remained negative. Silicon Motion reported record Q2 2026 revenue of $451.0Mn, up 32% sequentially and 127% y/y. Non-GAAP EPS of $2.43 beat the $2.13 consensus by 14%.

AI infrastructure is expanding beyond chips and data centers. RTX secured major multiyear missile contracts, AMD made an $8.2 billion bet on spatial intelligence, and NextEra advanced a $22 billion energy campus built around data-center demand. Together, these developments reveal a broader industrial investment cycle spanning defense, AI compute and power, highlighting how rising demand for technology, security and electricity is driving a new wave of long-term capital invest

I am downgrading the rating for Maxlinear from a STRONG BUY to a HOLD. The business is executing, but I think that the current valuation offers no base-case margin of safety.Q2 revenue rose 55% y/y to $168.8Mn, while Non-GAAP operating margin reached 22.3% and Q3 revenue guidance rose to $210Mn to $220Mn.

I am downgrading Diodes from a STRONG BUY to a HOLD. The operating evidence is better, but the old claim of asymmetric valuation upside no longer survives a disciplined present-value test. Q2 revenue rose 21.7% y/y to $445.5Mn, gross margin reached 33.1%, and adjusted EPS of $0.70 beat the $0.63 consensus by $0.07.

I downgrade ANET from Strong Buy to Hold. 2Q fy 2026 revenue reached $3.036Bn, up 37.7% year over year and 7.3% above consensus, while non-GAAP EPS of $1.02 beat by $0.13. The business is executing above the previous thesis, but the current $199.39 price leaves little base-case return.

The most important AI story this week was not a new benchmark or a single chip launch. It was a consumer product. Meta's Muse gave investors a tangible example of what mass market AI agents could look like, and that changed the conversation around the companies that distribute and supply the required compute.

The first phase of the artificial-intelligence trade was about securing enough compute. The second was about building the networks, data centres and power systems required to use that compute at scale. The next constraint may be smaller, less visible and just as important.