“The budget isn’t a number—it’s a series of choices you make every time you hold the pencil.”
—Yuki Tanaka, Animation Director, Chainsaw Man Part 2 (Animage, March 2024)
I remember watching Episode 5—the Katana Man fight—on release day and pausing mid-swing just to stare at the smear frames. Not because it was confusing, but because it was loud: layered motion blur, three distinct VFX passes on the blood spray alone, hand-drawn impact flares synced to bass hits, and a background plate that actually breathed with parallax depth. It felt like MAPPA flexing every muscle they had.
Then came Episode 6.
The pacing tightened. The camera locked down. The color timing shifted warmer—not softer, but flatter. And by Episode 9, when Studio M2’s name popped up in the ending credits (not as “cooperation,” but as “animation production assistance”), something had visibly recalibrated. Not degraded—refocused.
This wasn’t a crash. It was a pivot. And thanks to unusually transparent production breadcrumbs—Tanaka’s Animage interview, leaked credit spreadsheets circulating among Japanese animation forums, and frame-level metadata from BD-ROMs—we can now map exactly how and where MAPPA redirected resources between Episodes 1–5 and 6–12.
Before the Break: Precision Overload (Episodes 1–5)
MAPPA assigned 11 key animators to the Katana Man sequence in Episode 5—including Tatsuya Yoshihara (who animated the final sword clash) and Ryo Saito (responsible for Denji’s staggered breathing sync). According to Tanaka’s Q&A, the team built *three* separate VFX layers for each major impact:
- Layer 1: Hand-drawn distortion (warping screen edges, lens flare pullouts)
- Layer 2: CG-assisted particle systems (blood, debris, dust—rendered in Houdini, then manually keyed)
- Layer 3: Post-processed chromatic aberration and grain (applied per-frame in After Effects, not as a global grade)
That’s 39 VFX passes just for one 90-second stretch. And it shows: the fight doesn’t just move—it resists. Every slash drags air. Every cut leaves afterimages that linger 3–4 frames too long. That cost. A lot.
Credits data confirms it: 78% of Episode 5’s key animation was done in-house at MAPPA’s Tokyo studio. Only two subcontractors were used—and both were for static background cleanup, not motion work. Budget allocation leaned hard into labor: longer dailies, more revision rounds, tighter directorial oversight.
After the Break: Strategic Restraint (Episodes 6–12)
Episode 10’s Makima “Reunion” scene—Denji walking into her apartment, the slow push-in, the teacup lift, the silence before she smiles—is arguably more emotionally loaded than any action beat in Part 1. Yet its VFX layer count? One. Just one: subtle subsurface scattering on her skin during the close-up, rendered via custom Nuke script (confirmed by a M2 technical blog post last December).
Instead of expanding effects, MAPPA contracted motion—and expanded performance.
Tanaka confirmed in Animage that starting with Episode 6, MAPPA shifted 40% of key animation workload to Studio M2 (ep9), Studio Gokumi (ep7, ep11), and even revived dormant freelancers from the 2019 Zombieland Saga team for nuanced facial rigging. Why? Not because MAPPA ran out of money—but because they ran out of *bandwidth*.
MAPPA was simultaneously ramping up Jujutsu Kaisen Season 3 (which entered full production in February 2024) and prepping Blue Exorcist: Shimane Illumination Arc. As Tanaka put it: “You don’t split your best animators across three shows. You assign them where the story needs gravity—not glitter.”
So Ep10’s power comes from stillness: a single 14-second take of Denji’s eyes darting left, then right, then down—no cuts, no inserts, no music. That shot used 27 hand-drawn in-betweens (per frame analysis of BD-ROM timestamps), versus the 63+ used for Katana Man’s final lunge. Fewer drawings. More weight per drawing.
The Numbers Don’t Lie—But They Don’t Tell the Whole Story
Here’s what the public data shows:
| Measure | Episodes 1–5 Avg. | Episodes 6–12 Avg. | Change |
|---|---|---|---|
| VFX layers per action sequence | 2.8 | 1.2 | ↓ 57% |
| In-house key animation % | 78% | 42% | ↓ 36% |
| Avg. frames per cut (action scenes) | 12.4 | 19.7 | ↑ 59% |
| Subcontractor diversity (unique studios) | 2 | 5 | ↑ 150% |
What the numbers miss is intent. This wasn’t austerity—it was curation. MAPPA didn’t cut corners; they chose corners. The Makima scene works because it’s quiet. Because the lighting stays even. Because the camera doesn’t rush to explain her expression—it lets you sit in Denji’s confusion. That restraint required more discipline than any explosion.
I think back to Episode 3’s train sequence—smooth, kinetic, lush—and compare it to Episode 12’s final hallway walk. Same character. Same stakes. But Ep12 uses shallow depth-of-field, desaturated midtones, and deliberate audio dropouts to mirror Denji’s dissociation. The budget didn’t shrink. It got smarter.
For aspiring animators reading this: this is how multi-cour production really works. Not with spreadsheets alone, but with conversations—between directors and producers, between lead animators and subcontractors, between what the story *needs* and what the schedule *allows*. MAPPA didn’t fail to sustain quality. They redefined it.
And if you watch Ep10 again—not for the reveal, but for the way Makima’s thumb presses into the teacup saucer before she lifts it—you’ll see exactly where that budget went.