$COHR Coherent Corp: Scaling the AI Optical Infrastructure Bottleneck. Investment Thesis. New: 7/2/26. Coherent Corp is navigating a business model transition that goes beyond the typical product cycle upgrade — the shift from diversified photonics manufacturer to specialized AI data center optical infrastructure provider requires organizational focus, capital reallocation, and customer relationship development that are fundamentally different from managing a broad component portfolio. Jim Anderson's arrival as CEO represents the leadership reset that transition requires, and the strategic clarity around vertical integration and AI infrastructure focus is more coherent than the prior multi-segment positioning that diluted investor attention and capital allocation discipline. The NVIDIA partnership is the demand anchor that validates the strategic repositioning. AI cluster bandwidth requirements are the specific problem Coherent's high-speed transceivers address, and a direct relationship with the dominant AI accelerator supplier provides both near-term revenue visibility and longer-term product roadmap alignment. The same customer concentration risk that applies to other NVIDIA supply chain partners applies here — strategic partnership and customer dependency are two descriptions of the same relationship, and the distinction matters when pricing power and contract terms are negotiated over time. The 6-inch indium phosphide production expansion is the manufacturing investment that converts the strategic narrative into operational reality. Scaling InP production addresses both cost reduction and supply bottleneck relief simultaneously — two problems that have historically limited optical transceiver suppliers from capturing the margin upside that their technical differentiation should support. Whether the yield ramp on 6-inch wafers proceeds on the timeline and at the cost structure the financial projections assume is the manufacturing execution variable that determines how quickly the operating leverage thesis materializes. Co-packaged optics represents the next architectural shift that creates both opportunity and execution risk for Coherent. The transition from pluggable transceivers to co-packaged configurations requires different product formats, closer customer integration, and manufacturing capabilities that are still maturing across the industry. Being positioned for that transition is strategically important; executing through it while simultaneously scaling current-generation transceiver production is operationally demanding. Record bookings and revenue trajectory provide near-term confidence that the market repositioning is gaining commercial traction. The valuation premium reflects the operating leverage potential of a vertically integrated supplier with a bottleneck position in AI optical infrastructure — but operating leverage requires both volume scale and cost structure improvement to materialize simultaneously, and the capital expenditure requirements of the 6-inch InP expansion will pressure free cash flow during the period when the revenue ramp is still building toward the scale that justifies the investment. The high-stakes transition framing is accurate. Coherent has the technical assets, the manufacturing capability, and now the strategic focus to become a meaningful AI infrastructure compounder — but the complexity of executing a manufacturing technology transition, a leadership change, and a business model repositioning simultaneously leaves the path to that outcome narrower than the revenue momentum alone suggests.

$AOSL Strategic Transformation, GAI Power Potential, and Investment Thesis. Alpha and Omega Semiconductor Limited, is a small-cap firm transitioning from a simple component vendor to a provider of sophisticated power solutions. The company is currently navigating a financial trough characterized by declining margins and losses, yet it is aggressively funding research and development to capture growth in AI data centers, premium smartphones, and gaming. While leadership is technically proficient and deeply aligned with shareholders through significant insider ownership, they must still prove their ability to convert technical design wins into sustainable profitability. The investment profile is presented as a high-stakes operating leverage opportunity, where success depends on achieving a mid-term model of $1 billion in revenue and significantly improved gross margins. Ultimately, the texts frame the company as a "prove-it" story, balancing promising technological optionality against cyclical risks and historical execution challenges.

$CSCO Cisco’s Strategic Evolution: From Networking Legacy to AI Infrastructure Platform. Investment Thesis. Cisco is executing a meaningful pivot from legacy networking hardware toward AI infrastructure and software platform — a transition that carries genuine strategic logic given the company's installed base, silicon capabilities, and customer relationships across enterprise and hyperscaler markets. Silicon One and Acacia optics position Cisco credibly in the data center networking stack where AI workloads are concentrating capital, and the hyperscale order momentum reflects real demand pull rather than just repositioning narrative. The Splunk integration is the software layer that transforms the story from hardware refresh to platform compounding. Combining security and observability into a unified ecosystem creates cross-sell opportunities and recurring revenue potential that the legacy Cisco business model didn't support at scale. The integration is still in progress, and the degree to which Splunk's capabilities are genuinely unified with the core networking portfolio — versus operated as an adjacent business — will determine how much platform premium the combined entity deserves. The financial performance is strong, and capital returns remain consistent. The caution is entirely in the valuation. The stock's recent move has pulled forward a significant portion of the fundamental upside, and the current price embeds sustained AI infrastructure demand, successful Splunk integration, and continued share capture against well-positioned competitors. That's a demanding combination of assumptions to hold simultaneously, and any softening in hyperscaler capex or integration execution friction would pressure a multiple that has already re-rated aggressively. Competition from Arista and NVIDIA is the structural challenge that the incumbent advantage doesn't fully resolve. Arista has established deep hyperscaler relationships in AI networking and is not easily displaced. NVIDIA's networking ambitions — through InfiniBand and increasingly Ethernet — represent a different kind of competitive threat, one that comes bundled with GPU procurement decisions rather than competing on networking merit alone. Cisco's scale and breadth are advantages, but winning AI networking share requires product and relationship superiority, not just incumbency. Restructuring is the right operational response to the strategic pivot — reallocating resources toward AI readiness while managing the margin pressure that comes with a hardware-intensive growth phase is a necessary and credible move. Chuck Robbins has managed the transition with more strategic clarity than the prior cycle warranted, and the organizational changes signal that the prioritization is genuine rather than cosmetic. The reinvigorated incumbent framing is apt. Cisco has the assets, the customer base, and the technical capabilities to be a significant AI infrastructure beneficiary. The remaining question is whether the earnings compounding that the current valuation implies materializes at the pace the stock is pricing — and whether the competitive environment allows Cisco to hold the share gains it has recently captured.

We've acquired Rivanna, to bring full AI-native due diligence to the Rogo platform. Deal execution today is fragmented across data rooms, models, trackers, and inboxes. With Rivanna, and Rogo's Deal Room, deal teams can integrate natively with entire data rooms and get institutional-grade, cited analysis in minutes - not hours of manual review.

$VLO $MPC $PSX $DINO Primarily crude oil (including Russian Urals and Siberian Light grades, plus some Kazakh oil such as KEBCO blend) and refined petroleum products such as diesel/gasoil, fuel oil (mazut), gasoline/petrol, naphtha, and vacuum gasoil. The Sheskharis (also spelled Shekhsaris) terminal, operated by Transneft’s Chernomortransneft subsidiary in Novorossiysk, is a major Black Sea export complex and the endpoint of key Russian trunk pipelines. It has a design throughput capacity of around 75 million tons per year of oil and petroleum products, with large storage at the linked Grushovaya depot.18 •Crude oil forms the bulk of volumes (historically often in the range of roughly 500,000–700,000+ barrels per day under normal conditions). It handles Urals (the main grade), Siberian Light, and limited volumes of Kazakh oil routed via Russian pipelines.19 •Refined products arrive largely by pipeline and rail. These include diesel fuel/gasoil, fuel oil (also used for ship bunkering), gasoline, naphtha, and related middle distillates/vacuum gasoil. All types of petroleum products are exported from the facility. Berths at the terminal are configured for both large crude tankers and product carriers. Note that a separate nearby Caspian Pipeline Consortium (CPC) terminal handles the majority of Kazakh CPC Blend crude exports, while Sheskharis focuses on Transneft-handled Russian crude and products.

One of the most honest reads on enterprise AI returns⬇️ Jane Street just raised $14.6B in senior secured notes, the largest bond deal ever by a non-bank market maker. The 10-year tranche priced at 8.088%. Primarily a refi of a $5.5B floater and an $11B debt load, but tech infrastructure and trading expansion are stated uses. What makes this rare: Jane Street is one of the largest disclosed non-tech internal-GPU users, and one of the few with publicly priced debt against it: 4,032 liquid-cooled GPUs in Dallas, ~$6B committed to CoreWeave plus $1B of equity, early work on a self-financed 100-200MW build, targeting 10x into the hundreds of thousands. They sell no compute and no inference. The chips forecast asset prices. The output is directly monetized and the funding cost is observable. So, the bar is computable: At CRWV's 6-year depreciation life and 10% salvage, $100 of Jane Street capex must throw off $20.33 (20.3%) annually for six years plus $10 back at the end to set NPV at zero against 8%, 10-year money. Salvage assumptions barely move it. Every 10 points of recovery buys only ~136bp annually: 25% salvage requires an 18.3% cash yield, etc. So Jane Street is carrying 8% money against a ~20% break-even, even on modest salvage assumptions. Nobody with their track record takes that trade casually, a real signal about what monetized AI output can earn when someone can actually measure it. It's also a signal about what the bond market will underwrite: $14.6B, ten years out against a "six-year asset," no NVIDIA residual backstop, clearing on conservative salvage assumptions. Which leaves room for these rates to come down if and as the fundamentals materialize and for IG borrowers to try the same trade on even better terms.

$VLO $MPC $PSX $DINO $XOM $CVX Company-by-Company Management Commentary on Refining Margins From the Most Recent Earnings Call Synthesis: Margin Outlook The unanimous read across all six management teams is that the US refining margin environment remains well above mid-cycle through at least 3Q26 and into 2027, driven by geopolitically-induced supply disruption rather than a cyclical demand surge. The key risk to this view is a faster-than-expected return of Middle Eastern and Russian refining capacity, which all managements acknowledge but none are currently pricing into their base cases. The secondary risk is a concentrated turnaround wave in 2027 if deferred maintenance is pulled forward simultaneously across the industry. For now, the setup — tight inventories, constrained global supply, record US utilization, and structurally limited new clean-product capacity — is as constructive as the sector has seen in years. Valero Energy (VLO) Valero processed more crude while operating fewer refineries, lifting system utilization to ~98% in 2Q26, with the Gulf Coast and North Atlantic regions setting all-time throughput highs. (1) The USGC ULSD crack spread nearly tripled year-over-year to $43.52/bbl. Management expects refining margin capture to improve further in 3Q as crude premiums ease and market backwardation flattens, with the refining indicator up 31% in 3Q. (2) Management is structurally bullish on future mid-cycle margins relative to historical calculations, citing hydroskimming margins in Northwest Europe as the price-setter for product crack spreads, rising carbon credit costs, inflationary OpEx/CapEx pressures, and a favorable outlook on crude quality discounts — particularly for heavy sour crude. (3) Management stated a clear incentive to run significantly more heavy crude than in recent quarters, with heavy crude throughput capacity substantially expanded since the Port Arthur coker startup in 2023. (4) Quarterly capex guidance of $612mn is 77% above 2Q levels as Valero works to fully restart Port Arthur by year-end. (5) Marathon Petroleum (MPC) MPC ran at 94% system-wide utilization on ~3mn bpd throughput in 2Q26, with the Gulf Coast specifically at 100% utilization delivering $27 adjusted EBITDA/bbl. 3Q26 guidance calls for 3,005 mb/d total throughput, with Gulf Coast at 99%, Mid-Con at 87%, and West Coast at 96%. (6) R&M margin more than doubled to $36.3/bbl versus $17.6/bbl in 2Q25, with the blended US 3-2-1 crack spread also more than doubling year-over-year to $33.54/bbl. Record EPS of $17.73 beat the Street estimate of $13.73. (7) MPC management characterized the refining macro as constructive, citing >9mn bpd of planned and unplanned global capacity downtime — approximately 4mn bpd above historical norms — driven by Persian Gulf disruptions and Ukrainian attacks on Russian infrastructure. Management expects an enhanced mid-cycle environment through end of 2026 and into 2027, with US gasoline inventory well below the 5-year range and distillate at the bottom of its 5-year range. (8) MPC's Chief Commercial Officer noted: "We're all in max diesel mode," adding that "when you're in max mode of any one product for a significant period of time, it puts pressure on other products" — with impacts beginning to appear in the gasoline crack spread. (9) Product margins are expected to moderate in 3Q26 as outages normalize, though MPC expects them to remain well above mid-cycle levels, with WTI-linked cracks up >20% in 3Q and throughput guidance up ~2% sequentially. (10) Phillips 66 (PSX) PSX achieved ~96% refining utilization in 2Q26, above the industry average and consensus of ~94%, with an 86% clean product yield and a 98% capture rate. The company is guiding for mid-90% utilization going forward. (11) Realized refining margins reached $24.08/bbl in 2Q26, more than double the $11.25/bbl a year earlier. Central Corridor gross margin was $29.56/bbl and Gulf Coast was $24.25/bbl, with refining segment operating income of $3,086mn. (12) PSX management is bullish on refining cracks through Q3 and into 2027, citing ~8.4mn bpd of offline global refining capacity, low global product inventories, Chinese exports running ~50% below recent-year levels, and structurally widening WCS differentials from increasing Canadian and Venezuelan heavy crude supply. (13) PSX EVP of Marketing & Commercial stated on the earnings call that supply disruptions from the Iran war look set to weigh on markets for gasoline and diesel through 2027. (14) Of the projected 1.1mn bpd of new global refining capacity in 2026, PSX management noted only 350–400k bpd will be for clean products, as many new international facilities are petrochemical-focused. (15) PSX flagged that refiners are deferring some maintenance and repairs to capitalize on current high margins, which could lead to a more concentrated turnaround cycle and increased unplanned outage risk in the future. (16) HF Sinclair (DINO) DINO ran 639.7 MBPD at 94.3% utilization in 2Q26. Management guided 3Q26 crude runs of 590–620 MBPD, reflecting planned maintenance at El Dorado. (17) Adjusted refinery gross margin rose 57% year-over-year to $25.95/bbl, with the West region earning $30.57/bbl — well above mid-cycle and supporting substantial quarterly free cash flow. (18) DINO management expects global refining capacity outages to keep markets tight into 2027, supporting near-term margins, while acknowledging crack spreads will moderate as capacity returns. (19) Small-refinery exemption decisions expected in August could result in the return of up to $415mn in RINs for DINO, providing additional earnings upside. (20) Chevron (CVX) Chevron achieved record US refinery throughput of over 1mn bpd at 97% utilization in 2Q26. (21) Adjusted downstream earnings increased 449% year-over-year, with downstream profit of $4.87bn versus $737mn in 2Q25. Refining margins were a $1,140mn tailwind for the US downstream segment. (22) Chevron management stated on the 2Q26 call that products are tighter than crude globally, leading to widened cracks, with upward pressure on product pricing expected into 3Q and potentially beyond. Management noted demand destruction is not obvious at any significant scale. (23) 3Q26 downstream earnings will face a $175–225mn headwind from planned turnaround activity at El Segundo and Richmond refineries. (24) Chevron's California market position is strengthening structurally: its two California plants account for 49% of US capacity, and market share in the state is expected to increase following the closure of PSX's Los Angeles refinery in late 2025 and Valero's Benicia refinery in April 2026. (25) ExxonMobil (XOM) ExxonMobil reported 2Q26 adjusted EPS of $3.52/share, a slight miss versus consensus, partly due to refinery maintenance that prevented full capture of elevated gasoline, diesel, and jet fuel prices. Overall profit was $14.7bn. (26) Approximately 35% of ExxonMobil's global downstream capacity is located along the US Gulf Coast, giving it significant leverage to the current margin environment. Mid-cycle EBIT for XOM's refining assets is projected at $3.01bn, the highest among the group. (27) Key Readthroughs: Direction of U.S. Refining Margins 1. Structural Supply Disruption Is the Dominant Driver — and It's Durable The Iran war and Ukrainian strikes on Russian refining infrastructure have collectively removed a significant portion of global fuel-making capacity, leaving US refiners as the world's primary fuel supplier. (28) Global planned and unplanned refinery downtime exceeds 9mn bpd — approximately 4mn bpd above historical norms — with no near-term resolution in sight. (29) RIN-adjusted crack spreads for 3Q26 are well above mid-cycle across all US regions: North Atlantic $48.14/bbl, Gulf Coast $42.38/bbl, Mid-Con $40.21/bbl, and West Coast $46.01/bbl. (30) 2. Inventory Tightness Provides a Floor US commercial crude stockpiles have fallen to their lowest level since 2018, with Cushing inventories at their lowest since August 2014. (31) US diesel inventory is 11.6% below the 5-year average and gasoline is 6.9% below — both providing a structural floor under product crack spreads. (32) 3. New Capacity Additions Are Not a Near-Term Threat Of the ~1.1mn bpd of new global refining capacity expected in 2026, only 350–400k bpd is for clean transportation fuels; the remainder is petrochemical-focused. (33) Management across all companies (VLO, MPC, PSX, DINO) consistently flagged that new capacity additions are ramping slowly with low reliability, and that refinery rationalizations in the US (PSX LA, VLO Benicia) are removing supply from the tightest regional markets. (34) 4. "Max Diesel Mode" Is Creating Cross-Product Pressure With refiners optimizing for distillate to capture the widest cracks, gasoline crack spreads are beginning to feel secondary pressure. (35) This dynamic is worth monitoring: if gasoline cracks compress meaningfully while distillate remains elevated, capture rates for refiners with less flexible configurations could disappoint relative to headline crack spread levels. 5. Deferred Maintenance Is a Latent Risk PSX explicitly flagged that the industry is deferring maintenance to maximize margin capture, which could result in a more concentrated turnaround cycle and elevated unplanned outage risk in 2027. (36) This is a double-edged dynamic: near-term it supports utilization and margins, but it raises the probability of a more disruptive maintenance wave once the current margin environment normalizes. 6. Heavy Crude Differentials Are a Structural Tailwind Widening WCS differentials (now >$15/bbl) and incremental Venezuelan barrels flowing to the USGC are providing a feedstock cost advantage for complex refiners — particularly VLO, MPC, PSX, and DINO. (37) Valero's management sees an incentive to run significantly more heavy crude, with Port Arthur's coker restart expanding capacity to do so. (38)

Grok 4.6 is roughly the same performance as Fable 5 Max at an 85% discount. 80% cheaper for input tokens and 88% cheaper for output tokens. Pareto dominant. Grok 4.7 will be significantly better as is a much larger model with the Cursor and SpaceX data included in pretraining.